Treatment of Gestational Trophoblastic Disease in the 2020s

Abstract

Purpose of Review

This review demonstrates the evidence for new systemic anti-cancer treatments and how they integrate within conventional management for gestational trophoblastic neoplasia (GTN). We present the evidence on atypical placental site nodules, and how they incorporate within the GTN spectrum, as well as updates regarding GTN staging and follow up.

Recent Findings

First line treatment for GTN still lies in conventional chemotherapy, although the introduction of anti-PD1/PD-L1 immune checkpoint inhibitors has shown significant promise in management of relapsed disease, with responses reported in multiply relapsed choriocarcinomas as well as Epithelioid Trophoblastic Tumours and Placental Site Trophoblastic Tumours (ETT/PSTT). Following completion of treatment, ETT/PSTT still require life-long surveillance but for other GTN no recurrences have been detected after 7 years.

Summary

Checkpoint inhibitors are likely to play an increasing role in the future management of GTN management. Further refinement of prognostic factors to identify those most at risk of GTN recurrence is warranted so that surveillance can be focussed on those most at risk, whilst minimising unnecessary intervention for those at lower risk.

Introduction

Gestational trophoblastic disease (GTD) is a rare but important range of related conditions arising from the placenta. This includes the pre-malignant conditions of complete (CHM) and partial hydatidiform moles (PHM) and the malignant invasive mole, choriocarcinoma and rare placental site and epithelioid trophoblastic tumours (PSTT/ETT). CHM and PHM are the commonest form of GTD affecting around 1-2 per thousand pregnancies in most western countries (1). They typically present with bleeding in the first trimester of pregnancy and an abnormal ultrasound prompts evacuation of the uterine contents for pathological examination. The latter provides the diagnosis and results in subsequent monitoring of the pregnancy hormone human chorionic gonadotrophin (hCG). This ensures early recognition of malignant transformation through a plateaued or rising hCG level which occurs in about 16% and 1% of women with CHM and PHM, respectively. Fortunately, with appropriate treatment, cure rates for these patients with malignant transformation now approaches 100%. A key issue is how to achieve this rapidly and with the least toxicity.

CHM and PHM can give rise to all of the malignant forms of GTD, collectively referred to as gestational trophoblastic neoplasia (GTN). Conversely choriocarcinoma and PSTT/ETT can arise after any type of pregnancy. The frequency of choriocarcinoma after other types of pregnancy is estimated at about 1:50,000 and PSTT/ETT accounts for just 0.2% of all GTD (2). As these malignancies can present months or years after a pregnancy they are often detected late in their natural history. Fortunately, choriocarcinomas are highly chemosensitive and so even patients with advanced disease can achieve cure rates in excess of 94% (1, 3). PSTT/ETT tends to be more resistant to chemotherapy and so overall cure rates are lower at around 75-80% (2, 4).

This review will examine recent developments in the diagnosis and management of patients with GTN with a focus on refining patient stratification and therapies to reduce toxicity and improve survival outcomes in GTN. For details about the epidemiology, pathology and genetics of GTD as well as management of molar pregnancies the reader is referred elsewhere(1, 3, 5, 6).

Updates in the FIGO scoring system for GTN

All patients with post molar GTN and those with choriocarcinoma have their clinical, imaging and hCG data utilised to complete the FIGO scoring system to enable accurate selection of single verses multi-agent chemotherapy (Table 1). This scoring system does not apply to PSTT/ETT. It is important that the information used to populate the scoring system is current and up-to-date. The details of imaging investigations needed for suspected low risk verses high-risk disease have been extensively reviewed elsewhere (1, 7, 8).

Table 1. FIGO Scoring System for GTN.

Prognostic Factor	0	1	2	4
Age	<40	≥40	-	-
Previous pregnancy	Hydatidiform mole	Abortion	Term
Months since last pregnancy	Less than 4	4 to 6	7-12	>12
Pre-treatment hCG (IU/mL)	<103	103 to 104	>104 to 105	>105
Largest tumor size, including uterus	<3cm	3-4 cm	≥5cm	-
Site of spread	Lung	Spleen or kidney	GI Tract	Brain, liver
Number of metastases	-	1 to 4	5-8	>8
Previous Failed chemotherapy	-	-	Single Drug	≥2

Current Management options for low risk GTN

The indications for treating GTN have been extensively reviewed elsewhere (1, 6, 7). The most frequent reason is a plateaued or rising hCG level. A score of 0-6 indicates a low risk of disease developing resistance to single agent chemotherapy with either methotrexate and folinic acid (MTX/FA) or actinomycin D (ActD). In women who have completed their family and have no metastatic disease then a hysterectomy could be considered (9). For selected patients a further suction curettage may also prevent the need for chemotherapy but where the hCG is greater than 5000 IU/L and/or the disease is within the uterine wall rather than the cavity, repeated curettage is likely to fail and is not recommended (10).

It is well recognised that as the score rises the risk of resistance increases. Thus 90% of patients scoring 0-1 will be cured with single agent treatment but it is widely reported that only one third are cured with a score of 5-6. Importantly, if resistance develops to one single agent then the other can be employed with a high likelihood of subsequent success (11). At Charing Cross, we have used an hCG cut-off of initially 100 IU/L (12), then 300 IU/L (13) and subsequently 1000 IU/L (ISSTD 2017 meeting) below which we would use ActD in the event of MTX/FA failure and above which we employed EMA/CO. Most patients given ActD following MTX/FA achieved remission but the few failing were all salvaged with subsequent multi-agent treatment. We have therefore increased the hCG cut-off to 3000 IU/L to help spare more women the toxicity of EMA/CO chemotherapy.

Exploration of other single agents that are active, avoid hair loss and are less toxic than EMA/CO is of course attractive. In this regard, single agent carboplatin AUC 6 given 3 weekly has shown promise in the management of low risk disease failing MTX/FA in Sheffield but was not so successful in Brazil (14, 15). In the UK, we are currently piloting AUC 4 carboplatin 2 weekly for patients with an hCG between 3000-30,000 IU/L and only using EMA/CO for those with an hCG > 30,000 IU/L at the point of resistance to MTX/FA. However, until results are available, such use of carboplatin cannot be recommended for routine use elsewhere. In the UK low risk treatment continues until hCG normalisation and for 6 weeks consolidation thereafter, whilst other centres only give 4 weeks consolidation which may increase the risk of recurrence (16). In Brazil, some centres have tried to amend the MTX/FA regimen to avoid treatment at weekends by giving MTX and this appears to be similarly efficacious and may enable more flexibility (17).

Improvement on the sequential use of single agents has been attempted by combining both MTX and Act-D (18); however rates of vomiting were higher in the combination arm, and the need for IV (c.f. IM treatment only) demands greater resources. Thus, at present our centre still uses MTX/FA as initial therapy and in the event of resistance the use of either ActD or EMA/CO as second-line therapy has now been widely adopted (8).

“Intermediate Risk” Patients

Given the increased rates of resistance to single agents with a FIGO score of 5-6, some investigators have proposed giving all such patients multi-agent chemotherapy from the outset (6). Since all low risk patients who start single agent treatment are ultimately cured, this proposal is likely to expose more patients to unnecessary toxicity from EMA/CO or similar multi-agent therapies. The alternative is to seek new ways to improve identification of the FIGO 5-6 patients who become resistant to single-agent treatment and instead treat these selectively with multi-agent therapy. Several approaches have been proposed for use prior to starting chemotherapy including a uterine artery pulsatility index <1 (43), the presence of lung metastases(19) and choriocarcinoma with lung metastasis (19), but none are as yet validated. Alternatively, the rate of initial hCG decline can be used as an early predictor of single agent failure but this obviously does not avoid starting treatment with a single agent (20). Therefore, a new international collaboration is underway to try and improve the FIGO scoring system in low risk disease.

High and Ultra-high risk GTN

Women with a FIGO score of 7 and above all receive multi-agent therapy, typically with etoposide, methotrexate and actinomycin D alternating weekly with cyclophosphamide and vincristine, from the outset (EMA/CO) (7). Recent work has shown that those with a score of 13 or higher (deemed “ultra-high risk”) are at risk of early and late death. The former occurs within the first 4 weeks of commencing treatment mostly due to haemorrhage or metabolic complications from a high burden of disease (reviewed in (1)). Treatment with low dose etoposide 100mg/m2 and cisplatin 20 mg/m2 on days 1 and 2, repeated every week for 1–3 weeks has been shown to ameliorate this (21, 22). In some circumstances such as the presence of liver metastases, EP-EMA (note second day of EMA is omitted) rather than EMA-CO may help reduce late deaths from multi-drug resistant disease. Treatment with EMA/CO again continues until HCG normalisation, where the EMA is changed from a 2-day to 1-day regime, with a further 6-8 consolidation weeks depending on the presence of high-risk features such as brain or liver metastases. Late deaths due to multi-drug resistant disease are still an area of active investigation for new treatments (see below).

A recent retrospective analysis compared EMA rather than EMA-CO in GTN to see whether treatment intensity could be reduced without compromising efficacy (23). In this study no significant difference was seen between the two treatments; however the groups were not well matched precluding reliable comparison. Data on a 4-day MEA regimen (methotrexate, etoposide and actinomycin D) have also been published, although clear superiority over EMA-CO in terms of efficacy, toxicity or convenience has not been demonstrated (24) and the regimen requires more hospital visits. In China the FAEV regimen remains highly active and popular (6). It is unlikely that any randomised trials will be undertaken due to the paucity of cases and the large numbers of patients required to show small differences in efficacy.

GTN salvage approaches: high dose verses immunotherapy

Approximately 20% of high-risk patients will relapse following primary chemotherapy, though the majority (about 80%) can be salvaged with further treatment, with hCG levels at relapse shown to be a prognostic indicator (25). There are a number of different options for patients who relapse after initial treatment. For those who have previously been treated with single agent regimes, combination treatment with EMA-CO is first line. For those who have relapsed after EMA-CO, other options include EP/EMA (26), or the much less toxic paclitaxel and etoposide alternating two weekly with paclitaxel and cisplatin (27, 28). More recently we have been piloting escalated EP (Etoposide 500mg/m2 and Cisplatin 60mg/m2 for 1 day, 2 weekly) with early data presented in Toronto 2019 at the ISSTD meeting. High-dose chemotherapy (HDC) and peripheral blood stem cell support for GTN has also been used as salvage treatment with a recent retrospective analysis reporting remission rates of 41% (29), of which 20% were solely due to the high dose therapy. Interestingly, an hCG level >12IU/L before or after HDC, cancer stage and presence of metastases were indicative of adverse survival outcomes. However, toxicity from HDC is significant with 3/32 patients dying from the procedure and all being rendered likely infertile.

The significant advances in immunotherapy in recent years, alongside the fact that GTN (invasive mole and choriocarcinoma) strongly express PD-L1 (30, 31) has led to checkpoint inhibitor use in GTN. Pembrolizumab (anti-PD-1) has effectively induced complete responses in 75-80% of unresectable, chemo-resistant GTN including cases that had failed HDC (32–35). Currently, it is unclear how to select responding cases as all are PD-L1 positive and mutational burden is absent or extremely low in GTN. Detection of tumour-infiltrating lymphocytes and HLA-G expression may be important but much more work is required (32). The PD-L1 inhibitor avelumab has also shown efficacy in GTN, inducing complete serological response in approximately 53% of patients who had previously been treated with single agent chemotherapy (36). Anti PD-1/anti PD-L1 treatment is generally well tolerated with minimal toxicity reported in the majority of patients and presents a much less toxic alternative to HDC with autologous stem cell transplantation. It may well be that they play a role earlier in the treatment pathway with future GTN management.

A new member of the GTD spectrum: Atypical placental site nodules

Placental site nodules (PSN) and in particular those with atypical features (APSN) were always considered to be benign lesions. Today about 10-15% of APSN are recognised to either co-exist with or go on to develop into PSTT/ETT (37) and PSN has also been seen to co-associate with PSTT/ETT. Consequently, these lesions can no longer be ignored and should be included within the GTD spectrum. They typically present with vaginal bleeding months or years after a prior pregnancy and the diagnosis is made following a dilatation and curettage or hysteroscopic biopsy of the uterus. Affected patients need careful review to measure the serum hCG, ascertain a detailed pregnancy history and whether metastatic cancer is already present.

If the imaging (which should include a MRI pelvis, head and CT chest and abdomen, all with contrast) is negative and the patient has completed her family then the safest option is a hysterectomy. This will hopefully reveal no evidence of PSTT/ETT and the patient can be discharged. In the event there is an underlying PSTT/ETT then the question will arise as to the interval between the end of the causative pregnancy and the diagnosis of the cancer (see section on PSTT/ETT). However, if fertility preservation is desired in a patient with normal imaging then surveillance is needed even if the likely causative pregnancy was many years earlier. If the scans are abnormal then further investigation may well reveal an underlying cancer that is managed in the usual way for PSTT/ETT.

Current management of PSTT and ETT

Clinical presentations are protean and can include amenorrhoea, irregular vaginal bleeding or an incidental finding on a biopsy or hysterectomy specimen (2). PSTT/ETT differ from other forms of GTN as they are slower growing, produce less hCG for the volume of disease present, metastasize later and involve lymph nodes more frequently (2). As the disease is less chemosensitive, the use of single-agents is not appropriate and the FIGO scoring system is not used. Instead, stage-adapted treatment is determined by two independent poor prognostic factors: an interval of ≥48 months from the causative pregnancy and, most recently discovered, stage IV disease (2). Those presenting within 48 months as opposed to beyond this have a 10 year OS of 95% and 20%, respectively. Those with stage IV vs I disease irrespective of interval have an OS of 40% and 85%, respectively. These two independent prognostic factors also hold true in ETT (38). Consequently, to enable appropriately selected risk-adjusted treatment it is essential to have both a histological diagnosis providing material for genetic analysis to determine the causative pregnancy and also careful whole body imaging.

Stage I tumours (confined to the uterus) arising <48 months since the antecedent pregnancy are managed with a total abdominal hysterectomy including removal of any suspicious pelvic and retroperitoneal lymph nodes. Adjuvant systemic therapy is not required and 10-year survival is 100% (2). In contrast, if the PSTT/ETT originated from a pregnancy > 48 months previously, then such stage I patients and indeed and stage II-IV patients should be offered aggressive platinum-based chemotherapy including the option for experimental treatments such as HDC or immunotherapy. Giving EP/EMA followed by tandem HDC appears to have improved survival from zero to about 50% in these long interval poor risk stage I-IV patients (2). In addition, stage IV patients regardless of interval should receive this type of aggressive management. Residual masses after treatment should be excised wherever possible to confirm no active cancer remains. (4).

Although PSTTs and ETTs may not express PD-L1 as strongly as complete mole or choriocarcinomas (30), successes have been reported with anti-PD1 treatment (32, 35) and this is already playing an increasingly large role in the management of these relapsed chemo-resistant tumours. PSTT have also been shown to express EGFR and VEGF (39) which may provide further treatment avenues in the future. In the UK pembrolizumab is also being used in poor prognosis PSTT and ETT (interval of ≥4 years from the causative pregnancy) as adjuvant treatment after surgery. The long-term survival data are awaited.

Follow-up of GTN patients: when to stop?

Our practice is to follow up GTN patients for 10 years although data on 4201 low- and high-risk GTN patients showed no relapses after 7 years (40). The risk of relapse is highest in the first 12 months following treatment (72.7% or 86.4% of all relapses for low- and high-risk disease respectively (40)), and thus surveillance occurs with reducing frequency as time from treatment increases. For those patients wishing to conceive, they should ideally wait for 12 months to save confusion caused by a rising hCG. Those patients with invasive mole or choriocarcinoma generally require hCG surveillance only. PSTT/ETTs should have 6 monthly MRIs for the first 2-3 years ad annually thereafter at least for 5 years as a rising serum hCG is not so sensitive for identifying recurrence (2).

Follow up can be burdensome for patients as well as resource intensive, and thus identifying those at highest risk of relapse could permit a more targeted follow-up approach. A study involving 118 patients with relapsed GTN (invasive mole or choriocarcinoma) identified >12 month interval between antecedent pregnancy and chemotherapy and >14 weeks from first chemotherapy to achieving β-hCG normalisation as predictors of recurrence (41), and the presence of residual lung lesions (in either low-or high-risk patients was not indicative of relapse in a recent retrospective analysis (42). However as yet we do not routinely tailor our follow up based on their risk stratification.

Conclusion

GTD is a unique spectrum of diseases that arise from the placenta. Whilst the mainstay of their management remains in conventional chemotherapy, recent advances in immunotherapy have shown promise, particularly in the management of chemo-resistant disease. The rarity of GTN precludes large randomised trial data and as such patients should be managed in specialised centres that have greatest experience in managing patients with these rare but often treatable diseases.

Key Points.

• GTN comprises a rare but largely treatable range of malignancies arising from the placenta.

• Given their largely chemosensitive nature, the mainstay of first line management lies in conventional chemotherapy

• Checkpoint inhibitors have demonstrated promise in the management of relapsed disease, are significantly less toxic than high dose chemotherapy with bone marrow support, and are likely to be used more widely in future

• Further tailoring of follow up is required, but surveillance beyond 7 years after initial diagnosis is not necessary for most GTN

Proposed Double Bulleted.

• -You B, Bolze PA, Lotz JP, Massardier J, Gladieff L, Joly F, et al. Avelumab in Patients With Gestational Trophoblastic Tumors With Resistance to Single-Agent Chemotherapy: Cohort A of the TROPHIMMUN Phase II Trial. J Clin Oncol. 2020;38(27):3129-37. This is the first trial of immunotherapy used as second-line therapy to treat low or high risk GTN patients showing that the anti-PD-L1 monoclonal antibody avelumab was moderately effective in chemo resistant disease, and had a favourable safety profile compared to chemotherapy.
• -Balachandran K, Salawu A, Ghorani E, Kaur B, Sebire NJ, Short D, et al. When to stop human chorionic gonadotrophin (hCG) surveillance after treatment with chemotherapy for gestational trophoblastic neoplasia (GTN): A national analysis on over 4,000 patients. Gynecol Oncol. 2019;155(1):8-12. This is a key paper highlighting that no patients relapsed after 7 years, thus providing a framework for GTN follow up.
• -Froeling FEM, Ramaswami R, Papanastasopoulos P, Kaur B, Sebire NJ, Short D, et al Intensified therapies improve survival and identification of novel prognostic factors for placental-site and epithelioid trophoblastic tumours. Br J Cancer. 2019;120(6):587-94. This was the first paper to show that advanced stage was an independent adverse prognostic factor and showed that more intensive platinum based-therapy including high dose chemotherapy appears to be improving the survival of long-interval patients (50% vs 0% without this treatment).

Proposed Single Bulleted.

• -Ghorani E, Kaur B, Fisher RA, Short D, Joneborg U, Carlson JW, et al. Pembrolizumab is effective for drug-resistant gestational trophoblastic neoplasia. Lancet. 2017;390(10110):2343-5. The seminal paper showing that Pembrolizumb could induce sustained remission in three out of four multi-drug resistant GTN patients. An on-going multi-centre study of this therapy currently shows an 80% complete response rate.
• -Lok C, van Trommel N, Massuger L, Golfier F, Seckl M. Practical clinical guidelines of the EOTTD for treatment and referral of gestational trophoblastic disease. Eur J Cancer. 2020;130:228-40. This paper outlines an up to date consensus for referral and treatment of GTN from the European Organisation for Treatment of trophoblastic diseases (EOTTD) adopted by ESGO, ISSTD and GCIG
• -Bouchard-Fortier G, Ghorani E, Short D, Aguiar X, Harvey R, Unsworth N, et al. Following chemotherapy for gestational trophoblastic neoplasia, do residual lung lesions increase the risk of relapse? Gynecol Oncol. 2020;158(3):698-701. This paper demonstrated that residual lung lesions were not predictive of relapse following treatment for GTN.

Abbreviations

MTX/FA

Methotrexate, Folinic Acid

ActD

Actinomycin D

EMA-CO

Etoposide, Methotrexate and Actinomycin-D, Cyclophosphamide, Vincristine

FAEV

Floxuridine, Actinomycin-D, Etoposide, Vincristine

EP

Etoposide and Cisplatin

TE/TP

Paclitaxel, Etoposide/ Paclitaxel, Cisplatin

GTD

Gestational Trophoblastic Disease

GTN

Gestational Trophoblastic Neoplasia

CHM

Complete Hydatidiform Mole

PHM

Partial Hydatidiform Mole

PSTT

Placental Site Trophoblastic Tumour

ETT

Epithelioid Trophoblastic Tumour

APSN

Atypical Placental Site Nodule